Lupus nephritis (LN) affects up to 65% of systemic lupus erythematosus (SLE) patients and results in renal failure in up to 42% of patients after five years. Renal failure rates can vary up to eight fold over five years in aggressive LN despite use of the same treatment protocol. A more reliable means of determining prognosis is therefore required. The long-term objective of this proposal is to identify noninvasive urine protein markers of disease that are predictive of renal compromise. Previous studies in this area have resulted in the identification of a limited number of candidate proteins that correlate with disease type and glomerular inflammatory activity. To develop models of prognostic indicators in a heterogeneous disorder such as LN, the candidate protein approach has limited power. The Specific Aims of this proposal are to determine: 1) differentially expressed proteins in the urine of LN subjects (revealed by proteomic techniques and artificial neural network (ANN) technology) as surrogate markers of the WHO class, activity, and chronicity of glomerular lesions on renal biopsy and 2) the feasibility of a larger study designed to examine these urine protein markers as predictors of renal outcome in LN. To address these aims, urine from 1) LN subjects undergoing renal biopsy and 2) subjects with either LN or SLE without LN will be analyzed using proteomic techniques. All subjects will be followed prospectively for renal outcomes. Urine proteins will be concentrated and resolved by 2D-gel electrophoresis. Spots will be identified by location and removed for identification by matrix assisted laser desorption/ionization mass spectrometry and peptide mass fingerprinting. Data from these procedures will be analyzed using ANN modeling to determine surrogate urine protein markers of WHO class, activity, and chronicity of LN renal biopsies. ANN will be used to determine surrogate markers that predict compromise in renal function due to LN. These markers will be compared to renal biopsy and traditional laboratory and clinical indices of disease as predictors of renal outcome. Within the two years of this study, surrogate urine protein markers of biopsy findings should be identified, and the feasibility of a long-term outcome study should be determined. The ultimate results of this study could revolutionize the manner in which treatment for LN is determined. Due to the power of proteomic and ANN techniques, novel mediators of disease activity and damage should be identified and thus contribute to the development of more targeted therapies for LN.